Big data–driven observational evidence in Korea: challenges in interpreting non-randomized studies